CGAL users discussion list

[Andreas Fabri <>]

Hello,

In the pull request https://github.com/CGAL/cgal/pull/1991
I added the graph_traits and free functions so that now
all triangulation classes can be used with the BGL.

Concerning the edge weights, you might store them in
a boost::unordered_map<edge_descriptor,double>,
fill it as you like, wrap it in a boost associative
property map, and then pass it to the MST algorithm,
instead of redefining T2_edge_weight_map. The weight map
is a parameter of the MST algorithm.

best,

andreas

On 21/03/2017 03:34, Ch'Gans wrote:
> Hi there,
>
> I'm doing some experiment with CGAL Constrained Delaunay Triangulation
> and boost graph Minimum Spanning Tree.
>
> The constrained edges of CDT correspond to path that already exist,
> the non constrained edges of CDT are transient edges for the MST
> algorithm.
> To make sure that MST will use the constrained edges, i'm setting
> their associated 'weight' (length) to zero,  and for the
> non-constrained edges i'm setting their 'weight' to their squared
> length.
>
> I'm currently doing all of these manually, including 'transforming'
> the data structures between the steps of my 'pipeline'. I know it is
> not optimal, but at least i am able to get the whole pipeline working.
>
> I went through the chapter 'CGAL and the Boost Graph Library' (BTW big
> thumb up for the documentation!) and saw that CGAL provides all the
> bolts and nuts to bind CGAL with boost graph. For example, CGAL
> provide the edge weight map needed by MST, filters that allow to
> filter out infinite edges, ...
>
> I am now wondering if I could use a filter to hide/proxy the weight of
> the constrained edges or if i should use a different approach.
> Maybe I should redefine T2_edge_weight_map
> (CGAL/boost/graph/graph_traits_Triangulation_2.h) so that it takes a
> CDT as a ctor argument...
>
> Actually it looks like i cannot use graph_traits_Triangulation_2.h or
> graph_traits_Delaunay_Triangulation_2.h, i would need to write my own
> version for CDT, do I?
>
> Any clarification or point out are very welcome,
> Chris

--
Andreas Fabri, PhD
Chief Officer, GeometryFactory
Editor, The CGAL Project

phone: +33.492.954.912    skype: andreas.fabri